Arran

Load Packages

First I need to load up the packages I’ll need

library(sf)

Linking to GEOS 3.4.2, GDAL 2.1.2, proj.4 4.9.1

library(ggplot2) #development version!
## devtools::install_github("tidyverse/ggplot2")
library(tidyverse)

Loading tidyverse: tibble
Loading tidyverse: tidyr
Loading tidyverse: readr
Loading tidyverse: purrr
Loading tidyverse: dplyr
Conflicts with tidy packages -------------------------------------------------------
filter(): dplyr, stats
lag():    dplyr, stats

library(readr)
## Not sure about this bit
#library("tidyverse",lib.loc="/Library/Frameworks/R.framework/Versions/3.4/Resources/library")
library(cowplot)

Attaching package: ‘cowplot’

The following object is masked from ‘package:ggplot2’:

    ggsave

library(sp)

Import Data

Now I import my data. I filter for the Arran postcodes, (since Arran all begins ‘KA27’).

## Finding the Arran coordinates
library(dplyr)
allcoordinates <- read.csv("alldata/ukpostcodes.csv")
arrancoordinates <- filter(allcoordinates,substr(postcode,1,4)=="KA27")

Plot Coordinates

Now I plot these coordinates.

## Plotting the Arran coordinates
ggplot(data = arrancoordinates) +
  geom_point(mapping = aes(x = longitude, y = latitude)) +
  ggtitle("Arran Postcodes") +
  labs(title = "Arran Postcodes", x = "Longitude", y = "Latitude") +
  theme(plot.title = element_text(hjust = 0.5)) +
  theme_grey() +
  coord_map()

Now I create some plots. #Arran Borders

pcs <- read_sf("alldata/Scotland_pcs_2011")
#Print Post codes lists
arransubsect <- filter(pcs,substr(label,1,4)=="KA27")
arransubsect %>%
  ggplot() +
  geom_sf() +
  theme(axis.text.x=element_text(angle=45, hjust = 1)) +
  theme_grey()

After a little editing I can overlay the two.

simple.sf <- st_as_sf(arrancoordinates, coords=c('longitude','latitude'))
st_crs(simple.sf) <- 4326

simple.sf %>% ggplot() + geom_sf() +
theme_grey() 

arransubsect %>%
  ggplot() +
  geom_sf() +
  theme(axis.text.x=element_text(angle=45, hjust = 1)) +
  theme_grey() +
  geom_sf(data=simple.sf)

Shape files

Now I can load the SIMD data, starting with the shape files.

#Import SIMD data from http://www.gov.scot/Topics/Statistics/SIMD
#The "new data zone boundaries with SIMD16 ranks (zipped shapefile)"
#'2011 Data Zone boundaries'
DZBoundaries2016 <- read_sf("./alldata/SG_SIMD_2016")
#https://data.gov.uk/dataset/scottish-index-of-multiple-deprivation-simd-2012
#https://data.gov.uk/dataset/scottish-index-of-multiple-deprivation-simd-2012/resource/d6fa8924-83da-4e80-a560-4ef0477f230b
DZBoundaries2012 <- read_sf("./alldata/SG_SIMD_2012")
DZBoundaries2009 <- read_sf("./alldata/SG_SIMD_2009")
DZBoundaries2006 <- read_sf("./alldata/SG_SIMD_2006")
DZBoundaries2004 <- read_sf("./alldata/SG_SIMD_2004")

Load SIMD data

Then (having already downloaded it), I can load the SIMD data.

#Look at data from 2016
SIMD2016 <-read.csv("./alldata/00505244.csv")
SIMD20162 <-read_sf("./alldata/SG_SIMD_2016")
#Look at data from 2012
SIMD2012 <- readxl::read_excel("./alldata/SIMD2012/00410770.xls")
SIMD20122 <- readxl::read_excel("./alldata/SIMD2012/00416552.xls")
#Look at data from 2009
SIMD2009 <- readxl::read_excel("./alldata/SIMD2009/0096578.xls")
SIMD20092 <- readxl::read_excel("./alldata/SIMD2009/0097806.xls")
#Look at data from 2006
# 2009 data - SIMD2006 <- readxl::read_excel("./alldata/SIMD2006/0096578.xls")
SIMD20062 <- readxl::read_excel("./alldata/SIMD2006/0097880.xls")
#Look at data from 2004
SIMD2004 <- readxl::read_excel("./alldata/SIMD2004/0027003.xls")

Select Arran SIMD data

I have to choose the right columns manually in order to select the Arran data.

#Selecting ArranDZ2016
Arrandz <- c(4672,4666,4669,4671,4667,4668,4670)
#Health domain rank
#2016
arran2016 <- SIMD20162[Arrandz,]
#Find postcode look-up, KA27 postcodes. Find unique DZ. Find row positions.
#Selecting ArranDZ2012
Arrandz2012 <- c(4409,4372,4353,4352,4351,4350,4349)
#2012
arran2012 <- DZBoundaries2012[Arrandz2012,]
#2009
arran2009 <- DZBoundaries2009[Arrandz2012,]
#2006
arran2006 <- DZBoundaries2006[Arrandz2012,]
#2004
arran2004 <- DZBoundaries2004[Arrandz2012,]

The reason I’ve downloaded all the datazones shapefiles individually (three steps up), is because they change between 2016 and 2012.

arran20162 <- arran2016 %>%
  select(DataZone, geometry, Percentile)  %>%
  mutate(year="2016")
arran20122 <- arran2012 %>%
  select(DataZone, geometry, Percentile) %>%
  mutate(year="2012")
arran1612 <- rbind(arran20162,arran20122)

See the small difference.

arran1612 %>%
  ggplot() +
  geom_sf(aes(fill = DataZone)) +
  facet_wrap('year') +
  theme_grey() +
  theme(legend.position="none") +
  theme(axis.text.x=element_text(angle=45, hjust = 1))

Now I want to plot all the data, first I combine it all into one table. First I subselect the data I want from the appropriate columns.

arran20092 <- arran2009 %>%
  select(DataZone, geometry, Percentile) %>%
  mutate(year="2009")
arran20062 <- arran2006 %>%
  select(DataZone, geometry, Percentile) %>%
  mutate(year="2006")
arran20042 <- arran2004 %>%
  select(DataZone, geometry, Percentile) %>%
  mutate(year="2004")
arransimd <- rbind(arran20162,arran20122,arran20092,arran20062,arran20042)

Arran Percentile Plots

Now I plot the percentiles.

arransimd %>%
  ggplot() +
  geom_sf(aes(fill = Percentile)) +
  facet_wrap('year') +
  theme_grey() +
  theme(axis.text.x=element_blank(),
        axis.ticks.x=element_blank(), axis.text.y=element_blank(),
        axis.ticks.y=element_blank())

There we are. Not the SIMD health percentiles of Arran zones throughout SIMD history. And I’ve learned a little bit about graphics in R.

If I wanted to I could show the zones individually..

First I find the unique zones. (There are 14. 7 Zones 2016, 7 Zones pre-2016)

datazones <- unique(arransimd$DataZone)

I’ll have to find out a simpler way to do this but..

Pre-2016 Individual Zones

S01004409 <- filter(arransimd, DataZone=="S01004409")
S01004372 <- filter(arransimd, DataZone=="S01004372")
S01004353 <- filter(arransimd, DataZone=="S01004353")
S01004352 <- filter(arransimd, DataZone=="S01004352")
S01004351 <- filter(arransimd, DataZone=="S01004351")
S01004350 <- filter(arransimd, DataZone=="S01004350")
S01004349 <- filter(arransimd, DataZone=="S01004349")

pre2016list <- list(S01004409, S01004372, S01004353, S01004352, S01004351, S01004350, S01004349)

function1 <- function(argument) 
{
  argument %>%
  ggplot() +
  geom_sf(aes(fill = Percentile)) +
  facet_wrap('year') +
  theme_grey() +
  theme(axis.text.x=element_blank(),
        axis.ticks.x=element_blank())
}

function2 <- function(argument) 
{
  arransubsect %>%
  ggplot() +
  geom_sf() +
  theme_grey() +
  theme(axis.text.x=element_text(angle=45, hjust = 1)) +
  geom_sf(data= argument, aes(fill = DataZone))
}

function3 <- function(argument) 
{
  argument %>%
  ggplot() +
  geom_sf(data = arransubsect) +
  geom_sf(aes(fill = Percentile)) +
  facet_wrap('year') +
  theme_grey() +
  theme(axis.text.x=element_blank(),
        axis.ticks.x=element_blank())
}

function4 <- function(argument) 
{
  plot_grid((function1(argument)), (function2(argument)), labels = c("A", "B"))
}

lapply(pre2016list, function4)

[[1]]

[[2]]

[[3]]

[[4]]

[[5]]

[[6]]

[[7]]

2016

Now I apply the same functions to the post-2016 Zones

arransubsect <- filter(pcs,substr(label,1,4)=="KA27")
S01011177 <- filter(arransimd, DataZone=="S01011177")
S01011171 <- filter(arransimd, DataZone=="S01011171")
S01011174 <- filter(arransimd, DataZone=="S01011174")
S01011176 <- filter(arransimd, DataZone=="S01011176")
S01011172 <- filter(arransimd, DataZone=="S01011172")
S01011173 <- filter(arransimd, DataZone=="S01011173")
S01011175 <- filter(arransimd, DataZone=="S01011175")

post2016list <- list(S01011177, S01011171, S01011174, S01011176, S01011172, S01011173, S01011175)

lapply(pre2016list, function4)

[[1]]

[[2]]

[[3]]

[[4]]

[[5]]

[[6]]

[[7]]

Plot the percentiles as bar charts.

Pre-2016

arransimd2 <- dplyr::filter(arransimd, year < 2016)
arransimd2 %>%
ggplot(aes(x=year, y=Percentile)) +
  geom_bar(stat="identity") +
  facet_wrap('DataZone') +
  theme_grey() +
  labs(title = "Arran SIMD Datazones", x = "Year", y = "Percentile") +
  theme(plot.title = element_text(hjust = 0.5))

Post-2016

arransimd3 <- dplyr::filter(arransimd, year >= 2016)
arransimd3 %>%
ggplot(aes(x=year, y=Percentile)) +
  geom_bar(stat="identity") +
  facet_wrap('DataZone') +
  theme_grey() +
  labs(title = "Arran SIMD Datazones", x = "Year", y = "Percentile") +
  theme(plot.title = element_text(hjust = 0.5))

Now I want to overlay the postcodes for a particular shapefile, in this case by Datazone. To do this I’ve converted both the Arran coordinates and Arran (2016) shapefiles into Spatial Points/Polygons, converted them into a common CRS, and then compared them by using over().

exampleshapes <- sf:::as_Spatial(arran2016$geom)
examplepoints <- sf:::as_Spatial(simple.sf$geom)
examplepoints <- spTransform(examplepoints, CRS("+proj=longlat +datum=WGS84"))
exampleshapes <- spTransform(exampleshapes, CRS("+proj=longlat +datum=WGS84"))
namingdzpostcode <- over(exampleshapes, examplepoints, returnList = TRUE)

I can then take a member reference from the orginal postcode list, which gives me a selection of the rows in that DZ. For simplicity I’ve written this as a new function.

Unfortunately, I haven’t worked out how to coordinate the new ID with the original DZ names yet, so I have to select by using the appropriate ID for each DZ. #Function selecting simple.sf by DZ ID.

function6 <- function(argument) 
{
  simple.sf[namingdzpostcode[[argument]],]
}

I can then use the above, and plot over the appropriate DZ shapefile. e.g

Projecting the coordinate selections

function1(S01004372) +
  geom_sf(data=function6(2))

function2(S01004372) +
  geom_sf(data=function6(2))

function3(S01004372) +
  geom_sf(data=function6(2))

If I edit function 4 a little so that the geom_sf layer is a second argument then I can also use function 4.

function4.5 <- function(argument, argument2) 
{
  a <- function1(argument)
  b <- function2(argument) +
  geom_sf(data=function6(argument2))
  plot_grid(a, b, labels = c("A", "B"))
}

function4.5(S01004372,2)

I’ve also made another function to plot the DZ on it’s own with coordinates.

function5 <- function(argument, argument2) 
{
  argument %>%
  ggplot() +
  geom_sf() +
  theme_grey() +
  geom_sf(data=function6(argument2)) +
  theme(axis.text.x=element_blank(),
        axis.ticks.x=element_blank())
}

function5(S01004372,2)